You are on page 1of 15

G TECHNO

IN L
G

OG
R
EME

IE
S
THAT

RLD
WO
WI

LL
E
T H
CHANGE

THE TECHNOLOGY REVIEW TEN


hat if you had a crystal ball that foretold the future of technology? Imagine, for example, if
W you had known in 1990 just how big the Internet was going to be 10 years hence. Sorry, that
crystal ball doesn’t exist. But in this special issue of Technology Review, we offer you the next best
thing: the educated predictions of our editors (made in consultation with some of technology’s top
experts). We have chosen 10 emerging areas of technology that will soon have a profound impact
on the economy and on how we live and work. These advances span information technology,
biotechnology and nanotechnology—the core of TR coverage in every issue. All of these areas merit
special attention in the decade to come. In each area we’ve chosen to highlight one innovator who
exemplifies the potential and promise of the field. Keep this issue around and see how well our pre-
dictions hold up—even without the aid of that crystal ball. —The Editors

TECHNOLOGY REVIEW January/February 2001 97


Duke University’s Miguel Nicolelis handles
a robotic arm. Brain signals from an owl
monkey (seen on the monitors to the right)
control the arm’s movement.

98 TECHNOLOGY REVIEW January/February 2001


MIGUEL NICOLELIS
are showing that this idea is credible. At
Brain-Machine Emory University, neurologist Phillip
Kennedy has helped severely paralyzed
Interfaces people communicate via a brain implant
elle, a nocturnal owl monkey that allows them to move a cursor on a

B small enough to fit comfortably


in a coat pocket, blinks her out-
sized eyes as a technician plugs four con-
computer screen (see “Mind Over Mus-
cles,” TR March/April 2000). And implants
may also shed light on some of the brain’s
nectors into sockets installed in the top of unresolved mysteries. Nicolelis and other
her skull. In the next room, measure- neuroscientists still know relatively little
ments of the electrical signals from some about how the electrical and chemical
90 neurons in Belle’s brain pulse across a signals emitted by the brain’s millions of
computer screen. Recorded from four neurons let us perceive color and smell, or
separate areas of Belle’s cerebral cortex, give rise to the precise movements of
the signals provide a window into what Brazilian soccer players—whose photos
her brain is doing as she reaches to touch adorn the walls of the São Paolo native’s
one of four assigned buttons to earn her office. “We don’t have a finished model of
reward—a few drops of apple juice. how the brain works,” says Nicolelis. “All
Miguel Nicolelis, a Duke University neu- we have are first impressions.”
robiologist who is pioneering the use of
neural implants to study the brain, points Others in Brain-Machine Interfaces
proudly to the captured data on the com- Organization Project
puter monitor and says: “This readout is
Andy Schwartz Neural control of robotic arm
one of a kind in the world.” (Arizona State University)
The same might be said of Nicolelis,
John Donoghue Brain representation of
who is a leader in a competitive and high- (Brown University) movement
ly significant field. Only about a half-
dozen teams around the world are pur- Richard Andersen Improved neuroelectrode
(Caltech) systems
suing the same goals: gaining a better
understanding of how the mind works Phillip Kennedy, Roy Bakay Communication systems for
and then using that knowledge to build (Emory University) paralyzed patients
implant systems that would make brain
control of computers and other machines Nicolelis’ latest experiments, how-
possible. Nicolelis terms such systems ever, show that by tapping into multiple
“hybrid brain-machine interfaces” or neurons in different parts of the brain, it
HBMIs. Recently, working with the Lab- is possible to glean enough information
oratory for Human and Machine Haptics to get a general idea of what the brain is
at MIT, he scored an important first on up to. In Belle’s case, it’s enough infor-
the HBMI front, sending signals from mation to detect the monkey’s intention
individual neurons in Belle’s brain to a of making a specific movement a few
robot, which used the data to mimic the tenths of a second before it actually hap-
monkey’s arm movements in real time. pens. And it was Nicolelis’ team’s success
In the long run, Nicolelis predicts at reliably measuring tens of neurons
that HBMIs will allow human brains to simultaneously over many months—pre-
control artificial devices designed to viously a key technological barrier—that
restore lost sensory and motor functions. enabled the remarkable demonstration
Paralysis sufferers, for example, might with the robot arm.
gain control over a motorized wheel- Still, numerous stumbling blocks
chair or a prosthetic arm—perhaps even remain to be overcome before human
regain control over their own limbs. brains can interface reliably and comfort-
“Imagine,” says Nicolelis, “if someone ably with artificial devices, making mind-
could do for the brain what the pace- controlled prosthetic limbs or computers
maker did for the heart.” And, in much more than just lab curiosities. Among the
the same way that a musician grows to key challenges is developing electrode
feel that her instrument is a part of her devices and surgical methods that will
own body, Nicolelis believes the brain allow safe, long-term recording of neu-
will prove capable of readily assimilating ronal activities. Nicolelis says he’s begun
human-made devices. working with Duke’s biomedical engi-
Ongoing experiments in other labs neering department to develop a teleme-

Photograph by PAT R I C I A M C D O N O U G H TECHNOLOGY REVIEW January/February 2001 99


try chip that would collect and transmit IBM’s Cherie Kagan is making
data through the skull, without unwieldy transistors that could be far cheaper
sockets and cables. And this year Nicolelis and easier to fabricate than silicon
will become co-director of Duke’s new electronics. The reward: her own lab.
Center of Neuroengineering and Neuro-
computation, which will explore new com-
binations of computer science, chip design
and neuroscience. Nicolelis sees the effort
as part of an impending revolution that
could eventually make HBMIs as com-
monplace as Palm Pilots and spawn a
whole new industry—centered around
the brain. —Antonio Regalado

CHERIE KAGAN

Flexible Transistors
he implementation of pervasive

T computing—the spread of digi-


tal information throughout soci-
ety—will require electronics capable of
bringing information technology off the
desktop and out into the world (see
“Computing Goes Everywhere,” p. 52). To
digitize newspapers, product labels and
clothing, integrated circuits must be
cheap and flexible—a tough combination bridge and Pennsylvania State Universi- didn’t want to calculate [the speed],” says
for today’s silicon technology. Even the ty have made impressive progress, and Kagan. But she kept tweaking, and in less
cheapest form of silicon electronics— commercial products are nearing the than a year she had increased the mobil-
the cut-rate “amorphous” silicon used market. Last fall, for example, Philips ity of electric charges through her tran-
to drive laptop display screens—is too Research in Eindhoven, the Netherlands, sistors by four orders of magnitude—
pricey. What’s more, it’s difficult to incor- showed off the first prototype of a rudi- matching the speed of amorphous silicon
porate silicon electronics on bendable mentary display driven by polymer semi- and far exceeding most organic transis-
surfaces such as plastics. conductors. But there’s a catch: Organics tors. The results won her a staff position
Technology innovators are taking a are far slower than their silicon cousins. and her own lab at IBM.
couple of routes around these limits. Now, a 31-year-old materials scientist Kagan has since increased the speed
Some researchers are trying to reinvent at IBM, Cherie Kagan, may have opened by another 50 percent; further fine-tun-
amorphous silicon. Others have aban- the door to cheap, flexible electronics that ing, she believes, could provide at least
doned inorganic compounds like silicon pack the mojo needed to bring ubiquitous another doubling in acceleration. Not
computing closer. Her breakthrough? A only may the hybrids be far faster than
Others in Flexible Transistors compromise: transistors made from mate- amorphous silicon, they have a key
Organization Project rials that combine the charge-shuttling advantage over silicon-based electron-
power and speed of inorganics with the ics. Like some organic materials used to
Lucent/Bell Labs Organic circuits
(Murray Hill, NJ) affordability and flexibility of organics. make transistors, the hybrid materials
These hybrids were created by can be dissolved and printed onto paper
Richard Friend Organic light-emitting diodes
(University of Cambridge) chemist David Mitzi at IBM’s Thomas J. or plastic just like particles of ink. “I
Watson Research Center in Yorktown make my materials in a different lab and
Joseph Jacobson (MIT) Printed inorganics
Heights, N.Y. By the time Kagan arrived carry them over and add some liquid
Thomas Jackson Organic transistors at Watson in 1998 following a stint at Bell and spin them on,” says Kagan. “It’s not
(Penn State) Labs (she earned a PhD from MIT in very sophisticated, which is sort of the
1996), Mitzi had already shown that his goal, right? You really want it to be cheap.”
to develop transistors based on organic materials possessed intriguing electron- Thomas Jackson, a transistor expert
(carbon-based) molecules or polymers. ic properties. Kagan had a hunch they at Penn State who is developing organic
These organic electronics are inexpensive might make good transistors. But she circuits, says Kagan’s “fledgling results”
to manufacture and compatible with needed quick results; she’d been hired could pave the way for fast yet flexible
plastic substrates. Indeed, research teams as a postdoc—a limited-time offer. electronics. Jackson credits Kagan with
at places such as Lucent Technologies’ At the outset, the transistors flipped seizing the opportunity. “Not only does
Bell Labs, England’s University of Cam- on and off sluggishly. “The first times, I she have her own pocket of competence,

100 TECHNOLOGY REVIEW January/February 2001 Photograph by L A R R Y B U S A C C A


but she’s able to look around and see U S A M A FAY YA D repairs. The idea, he says, was to enable
exciting possibilities and then bring things any GM service technician to ask the
together. IBM has been working on these Data Mining database a question based on the model
sorts of materials for some time, but it ello again, Sidney P. Manyclicks. of car, the engine capacity, and so on, and

took the energy and enthusiasm and
vision and perspective of Cherie to trans-
late that into a thin-film transistor.”
H We have recommendations for
you. Customers who bought this
title also bought...”
get a quick, appropriate response. Sounds
straightforward. But, recalls Fayyad, “there
were hundreds of millions of records—no
The transistors could compete with Intrusive? A touch of personal atten- human being could go through it all.” The
organic electronics in a variety of appli- tion in the sterile world of e-shopping? pattern recognition algorithm he devised
cations, such as radio-frequency product Both, perhaps—but definitely a tour de to solve that problem became his 1991
ID tags. And then there’s the $20 billion- force of database technology. Conven-
per-year market for flat-panel video dis- tional databases sort though a few Others in Data Mining
plays, where the speed of Kagan’s tran- megabytes of structured data to find Organization Project
sistors could really make a difference. answers to specific queries. But compil-
Howard Wactlar Search very large video collections
Quicker circuits would deliver sharper ing a simple recommendation list requires (Carnegie Mellon)
displays than those driven by amorphous a system that can burrow through giga-
silicon at a fraction of the cost. That bytes of Web site visitor logs in search of Marti Hearst (University Automated discovery of new information
of California, Berkeley) from large text collections
would open the door to affordable wall- patterns no one can anticipate in advance.
sized displays or high-quality displays Welcome to data mining, also known Nokia Research Center Finding recurrent episodes in event
(Helsinki, Finland) sequence data
that pop out of your pen. If all goes well, as knowledge discovery in databases
the materials could be used in cheap, (KDD): the rapidly emerging technology Raghu Ramakrishnan Visual exploration of data on the Web
flexible displays within five years. that lies behind the personalized Web (University of Wisconsin)
Of course, bright displays that fit in and much else besides. The emphasis
your pocket will require portable power, here is on “emerging,” says Usama Fayyad, doctoral dissertation, which is still among
and that has Kagan pondering her next who should know: data mining didn’t the most cited publications in the data-
research challenge: cheap, flexible mate- exist as a field until he helped pioneer it. mining field.
rials for solar cells to liberate pervasive In 1987, the Tunisian-born comput- Data mining proved to have surpris-
computing from bulky batteries. “You er scientist was a graduate student at the ingly broad application. Fayyad left Michi-
aren’t going to want to carry a battery University of Michigan. He had taken a gan for NASA’s Jet Propulsion Laboratory,
around with your lightweight flexible summer job with General Motors, which where he applied his techniques to astro-
display,” she says. —Peter Fairley was compiling a huge database on car nomical research. In particular, his algo-

DigiMine’s Usama Fayyad devises algorithms that detect meaningful patterns in massive collections of information.

Photograph by K A R E N M O S K O W I T Z TECHNOLOGY REVIEW January/February 2001 101


ithm helped in automatically determining tions of free-form text documents. The
which of some two billion observed celes- results are still preliminary, as various labs
tial objects were stars and which were experiment with natural-language pro-
galaxies. The tool also helped find volca- cessing, statistical word counts and other
noes on Venus from the huge number of techniques. But the University of Califor-
radar images being transmitted from space nia at Berkeley’s LINDI system, to take one
probes. A geologist could retrieve the example, has already been used to help
image of a previously identified volcano; geneticists search the biomedical literature
the computer would then examine the and produce plausible hypotheses for the
picture for patterns and search through function of newly discovered genes.
other images for similar patterns. That Another hot area, says Fayyad, is
worked so well, Fayyad says, that “pretty “video mining”: using a combination of
soon the military intelligence people were speech recognition, image understanding
all over us, wanting to use it. And so were and natural-language processing tech-
doctors, who wanted to do automatic niques to open up the world’s vast video
searches of radiology images.” In 1995, in archives to efficient computer searching.
response to this widening interest, Fayyad For instance, when Carnegie Mellon Uni-
and his colleagues planned a full-scale versity’s Informedia II system is given
international conference on KDD. The an archive of, say, CNN news clips, it
conference drew about 500 participants, produces a computer-searchable index
more than double what had been expect- by automatically dividing each clip into
ed. (KDD 2000 drew 950.) individual scenes accompanied by tran-
By this time, with the Internet gush- scripts and headlines.
ing information onto everyone’s desk- Fayyad hopes that ultimately the
top, the urgency for data mining was techniques of data mining will become so
becoming evident in the corporate world. successful and so thoroughly integrated
IBM and other industry giants sensed a into standard database systems that they
market—and wanted in. Microsoft set will no longer be thought of as exotic.
its sights on Fayyad and enticed him to “People will just assume that their data-
join the company’s research labs. “They base software will do what they need.”
suggested that I take a look at databases —M. Mitchell Waldrop
in the corporate world,” says Fayyad. “It
was pretty sad. In many companies, the
RANJIT SINGH
‘data warehouses’ were actually ‘data
tombs’: the data went in and were never
looked at again.” Fayyad joined Microsoft
Digital Rights
in 1996 and organized a new research Management
group in data mining. “We looked at new itting in his office in McLean,
algorithms for scaling up to very large
databases—gigabytes or larger,” he says.
By decade’s end, Fayyad had caught
S Va., Ranjit Singh is at ground
zero of what may be the
biggest—and bloodiest—of the many
ContentGuard’s Ranjit Singh manages
access to online digital property.

the entrepreneurial bug sweeping through battles that will shape the Internet during number of startups. “The Internet,”
computer science labs. “I realized that the 21st century’s initial decade. The bat- Singh continues, “allows perfect repro-
even the organizations that loved the tle lines are sharply drawn. On one side duction of digital content and totally
idea of data mining were having trouble are owners of intellectual property, or frictionless distribution.” A few mouse
just maintaining their data.” What they content: books, music, video, photo- clicks sends a work to millions of users,
needed, he reasoned, was a company to graphic images and more. On the other but the creators and owners of the con-
host their databases for them, and provide are Internet users—think Napster—who tent won’t necessarily collect dime one
data-mining services on top of that. The want content to be freely distributed. (see “Your Work Is Mine!” TR Novem-
result was digiMine, a Kirkland, Wash., And then there is Singh, president of ber/December 2000).
startup that opened for business in March ContentGuard, a company that spun Ouch! You can bet the pain felt by
2000 with Fayyad as CEO. out of research at Xerox’s Palo Alto content owners who see their stuff flying
And the future of data-mining tech- Research Center (PARC) on a mission to everywhere via the Net will translate into
nology? Wide open, says Fayyad—espe- commercialize content protection in a action. Which is what Singh and Con-
cially as researchers begin to move beyond wired world. “The Internet changes tentGuard are about. Digital rights man-
the field’s original focus on highly struc- everything,” says Singh, 48, an England- agement, or DRM, is “the catalyst for a
tured, relational databases. One very hot born technology manager whose resume revolution in e-content,” says Singh.
area is “text data mining”: extracting unex- glitters with senior positions at Xerox, “DRM will allow content owners to get
pected relationships from huge collec- Citibank and Digital Equipment plus a much wider and deeper distribution than

102 TECHNOLOGY REVIEW January/February 2001 Photograph by C H R I S H A R T L O V E


able to buy the content they want “with-
out needing special viewers or down-
loads and without putting the user
through hoops,” he argues. To resolve
that, ContentGuard has forged multiple
partnerships with digital standard-bear-
ers such as Microsoft and Adobe Sys-
tems, and has extended its technology
so that it applies across many media,
including books, music and video.
Captivating as the possibilities of
DRM are, it is still in its early days. Says
John Schwarz, CEO of Reciprocal, Inc., a
distributor of ContentGuard and other
DRM solutions: “We are probably a year
or so away from seeing broad adoption of
DRM by the marketplace.”
Some analysts are more skeptical:

Others in Digital Rights


Organization Project
InterTrust Technologies Develops peer-to-peer
(Santa Clara, Calif.) distributed DRM technology

Reciprocal DRM clearinghouse


(New York)

Digimarc Watermarking to embed an


(Tualatin, Ore.) imperceptible code

Alchemedia “Clever Content” platform


(San Francisco) safeguards digital content

“I’m not convinced content can be pro-


tected in the Internet era,” says Eric
Scheirer, who tracks DRM issues for For-
rester Research. “People want flexible
access to content.” Proof is Napster, of
course, which represents a phenomenon
Scheirer calls “unstoppable.” Even if Nap-
ster is put out of business by the courts,
he predicts that the frictionless distribu-
tion of digital content among the millions
ever before,” he maintains. “You can see So why isn’t DRM ubiquitous? Two of Internet users will live on.
who is passing your content to whom.” reasons. First, content owners are in the But Singh is betting heavily that
Stripped to its essence, DRM—as midst of a hard rethink about both pric- DRM will prevail and, ironically, he
provided by ContentGuard and a number ing and distribution. Suddenly they are thanks Napster. “Napster turned this
of competitors—amounts to an encryp- wrestling with issues of how to price whole issue into a CEO-level question.
tion scheme with a built-in e-business three listens to a song, say, or a download The very highest corporate officers now
cash register. Content is encoded, and to of a low-resolution image that cannot are looking into content management
get the key a user needs to do some- be forwarded to others. “Content owners issues, and they want to protect their
thing—maybe paying money, maybe pro- now are trying out different economic property.”
viding an e-mail address. DRM providers models for valuing content,” says Singh, That, says Singh, augurs wider use of
deliver the protection tools; it is up to con- whose company provides DRM tools to, DRM. “Here’s the virtual cycle you will
tent owners to set the conditions. Con- among others, John Wiley & Sons and see: The more content a business puts
tentGuard uses a “multiple key” approach Houghton Mifflin. “DRM opens many online, the faster it will want to put still
to content protection; anyone who possibilities,” he adds. more content up, because it will see the
received bootleg content would have to The second issue may be the more economic benefits and users will see the
crack into it all over again. Thus, Singh nettlesome: “The user experience has to benefits of gaining access to more con-
explains, “even if a hacker cracks into a hide the complexity of the protection tent. That’s why we are seeing an explo-
piece of content, he cannot distribute it.” technologies,” says Singh. Users need to be sion here.” —Robert McGarvey

TECHNOLOGY REVIEW January/February 2001 103


J O S E P H AT I C K

Biometrics
n one sense, the field of biomet-

I rics—identifying individuals by
specific biological traits—has
already emerged. Large companies use
fingerprint sensors for logging on to cor-
porate networks, state driver’s license
authorities employ face recognition for
capturing and storing digital pho-
tographs, and the first iris-scan-protect-
ed ATM in the nation was introduced in
Texas in May 1999. Yet consumers have
been reluctant to adopt the technology,
and so far, it remains largely relegated to
military and government applications.
But the emergence of another tech-
nology—the wireless Web—could soon
change all that, according to Joseph Atick,
president and CEO of Visionics, one of the
leaders in face recognition technology.
“Personal digital assistants (PDAs) and cell
phones are becoming our portal to the
Visionics’ Joseph Atick sees the wireless
world, our transaction devices, our ID
Web as key to widespread consumer
and maybe one day our passport,” says
adoption of biometric technologies.
Atick. But entrusting these small gadgets
with so much of our personal and finan-
cial information carries with it a great skeptical of Atick’s vision of a biometric- to this process,” says Atick. In 1994, he and
risk.“It is this need for security,” Atick says, enabled wireless Web can’t deny his inge- colleagues Paul Griffin and Norman
“that is going to drive biometrics.” nuity and ambition. At the age of 15, Redlich founded Visionics.
And while the need for security is while living in Israel, Atick dropped out Based in Jersey City, N.J., Visionics
pushing the demand for biometric sys- of school to write a 600-page physics develops and markets pattern-recognition
tems, other technology developments— textbook entitled Introduction to Modern software called FaceIt. In contrast to the
increased bandwidth, new cell phones Physics. “I was bored in school. I wanted main competing technology, which relies
and handheld computers equipped with to show the establishment I was serious on data from the entire face, FaceIt veri-
digital cameras—will create an infra- about my interests,” says Atick. “This fies a person’s identity based on a set of
structure capable of putting biometrics book was my ticket to grad school.” four facial features that are unique to
into the hands of consumers. Visionics is Remarkably, Stanford University accept- the individual and unaffected by the pres-
taking advantage of this combination of ed him at 16 into its graduate program, ence of facial hair or changes in expres-
need and infrastructure by developing where he earned his master’s degree in sion. In the past few years, the system has
tools to enable people to authenticate physics and PhD in mathematical physics. found success fighting crime in England
any transaction they make over the wire- After graduation, Atick applied his and election fraud in Mexico.
less Web using their own faces. math skills to the study of the human In October, the company signed a
Even those in the industry who are mind. While heading the Computational merger agreement with Digital Biomet-
and Neuroscience Laboratory at Rocke- rics, a Minnetonka, Minn.-based bio-
feller University, he sought to understand metric systems engineering firm. Togeth-
Others in Biometrics
how the brain processes the abundance of er they plan to build the first line of
Organization Project visual information thrown at it by the “biometric network appliances”—com-
Visage Technology Face recognition environment. He and his colleagues dis- puters hooked to the Net with the capac-
(Littleton, Mass.) covered that the brain deals with visual ity to store and search large databases of
Iridian Technologies Iris recognition information much as computer algo- facial or other biometric information.
(Marlton, N.J.) rithms compress files. Because everyone The appliances, containing customers’
DigitalPersona Fingerprint recognition has two eyes, a nose and lips, the brain identification data, can then receive
(Redwood City, Calif.) extracts only those features that typically queries from companies wanting to
Cyber-SIGN Dynamic signature show deviations from the norm, such as authenticate e-transactions. And while
(San Jose, Calif.) verification the bridge of the nose or the upper cheek- consumers will be able to access the sys-
bones. The rest it fills in.“We soon realized tem from a cell phone, PDA or desktop
T-NETIX (Englewood, Colo.) Voice recognition
there was tremendous commercial value computer, Atick expects handheld devices

106 TECHNOLOGY REVIEW January/February 2001 Photograph by N I C K C A R D I L L I C C H I O


to be the biggest market. Visionics is also the solar system as articulate as HAL. industry giants such as IBM and
working with companies in Japan and But maybe it wasn’t that far off. HAL’s Microsoft, which see not only immediate
Europe to have FaceIt software installed on modern-day counterparts are catching applications for computer users who
new Web-ready mobile devices so con- up fast (sans the homicidal tendencies, need to keep their hands and eyes free but
sumers can capture their own faces and one hopes). Already we have commercial also the rapid evolution of speech-
submit encrypted versions over the Net. speech recognition software that can take enabled “intelligent environments.” The
Is that it for PINs and passwords? dictation, speech generation equipment day is coming when every object big
Atick predicts it will still be two to three that can give mute people voices and enough to hold a chip actually has one.
years before PDA- and cell-phone-wield- software that can “understand” a plain-
ing consumers are likely to use biometrics English query well enough to extract the Others in Language Processing
instead. And as futuristic as his vision is, right answers from a database. Organization Project
he is really striving toward something a Emerging from the laboratories,
Victor Zue (MIT Laboratory Conversational interfaces
bit old-fashioned. “Essentially, we are moreover, is a new generation of inter- for Computer Science)
bringing back an old element of human faces that will allow us to engage com-
Alexander I. Rudnicky Verbal interaction with small
commerce,” says Atick—restoring the puters in extended conversation—an (Carnegie Mellon) computers
confidence that comes with doing busi- activity that requires a dauntingly com-
ness face to face. —Alexandra Stikeman plex integration of speech recognition, Ronald A. Cole Domain-specific conversational
(University of Colorado) systems
natural-language understanding, dis-
course analysis, world knowledge, rea- BBN Technologies Dialog agent
KAREN JENSEN (Cambridge, Mass.)
soning ability and speech generation. It’s
Natural Language true that the existing prototypes can only
talk about such well-defined topics as We’d better be able to talk to these objects
Processing weather forecasts (MIT’s Jupiter), or local because very few of them will have room
he 1968 film 2001: A Space movie schedules (Carnegie Mellon’s for a keyboard.

T Odyssey gave us a vision of the


millennium based on the tech-
nological predictions of the day. One
Movieline). But the Defense Advanced
Research Projects Agency (DARPA) is
working on wide-ranging conversation-
Getting there will be a huge chal-
lenge—but that’s exactly what attracts
investigators like Karen Jensen, the gung-
result: HAL 9000, a computer that con- al interfaces that will ultimately include ho chief of the Natural Language Pro-
versed easily with its shipmates like any pointing, gesturing and other forms of cessing group at Microsoft Research. Says
other crew member. The timing was off: visual communication as well. Jensen: “I can’t imagine anything that
In the real 2001, there’s not a computer in Parallel efforts are under way at would be more thrilling, or carry more
potential for the future, than to make it
possible for us to truly interact with our
computers. That would be so exciting!”
Such declarations are typical of
Jensen, who at 62 remains as exuberant
about technology’s promise as any teenag-
er—and just as ready to keep hacker’s
hours. Indeed, Jensen was one of the first
people Microsoft hired when it opened its
research lab in 1991. Along with col-
leagues Stephen Richardson and George
Heidorn, she arrived at the Redmond,
Wash., campus from IBM’s Thomas J.
Watson Research Center, where they had
worked on some of the earliest gram-
mar-checking software, and immediate-
ly started building a group that now
numbers some 40 people.
In Redmond, Jensen and her col-
leagues soon found themselves con-
tributing to the natural-language query
interface for Microsoft’s Encarta ency-
clopedia and to the grammar checker
that first appeared in Word 97. And now,
Microsoft Research’s Karen Jensen is
she says, they’ve begun to focus all their
heading an effort to give machines the
efforts on a unique technology known as
ability to grok human language.
MindNet. MindNet is a system for auto-
matically extracting a massively hyper-

Photograph by K A R E N M O S K O W I T Z TECHNOLOGY REVIEW January/February 2001 107


linked web of concepts from, say, a J O H N J OA N N O P O U LO S ductors are to electrons, offering an excel-
standard dictionary. If a dictionary lent medium for controlling the flow of
defines “motorist” as “a person who dri- Microphotonics light. Like the doorman of an exclusive
ves a car,” for example, MindNet will use ight bounces off the small yellow club, the crystals admit or reflect specif-
its automatic parsing technology to find
the definition’s underlying logical struc-
ture, identifying “motorist” as a kind of
L square that MIT physics profes-
sor John Joannopoulos is show-
ing off. It looks like a scrap of metal,
ic photons depending on their wave-
length and the design of the crystal. In the
1990s, Joannopoulos suggested that
person, and “drives” as a verb taking something a child might pick up as a defects in the crystals’ regular structure
motorist as a subject and car as an object. plaything. But it isn’t a toy, and it isn’t could bribe the doorman, providing an
The result is a conceptual network that metal. Made of a few ultrathin layers of effective and efficient method to trap the
ties together all of human understanding non-conducting material, this photonic light or route it through the crystal.
in words, says Jensen. crystal is the latest in a series of materials Since then, Joannopoulos has been a
The very act of putting this concep- that reflect various wavelengths of light pioneer in the field, writing the definitive
tual network into a computer takes the almost perfectly. Photonic crystals are book on the subject in 1995: Photonic
machine a long way toward “under- on the cutting edge of microphotonics: Crystals: Molding the Flow of Light. “That’s
standing” natural language. For example, technologies for directing light on a the way John thinks about it,” says MIT
to figure out that “Please arrange for a microscopic scale that will make a major materials scientist and collaborator Edwin
meeting with John at 11 o’clock” means impact on telecommunications. Thomas. “Molding the flow of light, by
the same thing as “Make an appointment In the short term, microphotonics confining light and figuring out ways to
with John at 11,” the computer simply has could break up the logjam caused by the make light do his bidding—bend, go
rocky union of fiber optics and electron- straight, split, come back together—in the
Others in Microphotonics ic switching in the telecommunications smallest possible space.”
Organization Project backbone. Photons barreling through the Joannopoulos’ group has produced
network’s optical core run into bottle- several firsts. They explained how crys-
Eli Yablonovitch Photonic crystals for optical
(UCLA) and radio frequencies necks when they must be converted into tal filters could pick out specific streams
the much slower streams of electrons of light from the flood of beams in wave-
Susumu Noda Optical integrated circuits
(Kyoto University, Japan) that are handled by electronic switches length division multiplexing, or WDM,
and routers. To keep up with the Inter- a technology used to increase the
Axel Scherer Optical switches, waveguides
(Caltech) and lasers
net’s exploding need for bandwidth, tech- amount of data carried per fiber (see
nologists want to replace electronic “Wavelength Division Multiplexing,” TR
Nanovation Technologies Integrated devices for telecom switches with faster, miniature optical March/April 1999). The lab’s work on
(Miami)
devices, a transition that is already under two-dimensional photonic crystals set
Clarendon Photonics Filters for WDM way (see “The Microphotonics Revolu- the stage for the world’s smallest laser
(Boston)
tion,” TR July/August 2000). and electromagnetic cavity, key com-
Because of the large payoff—a much ponents in building integrated optical
to parse the two sentences and show that faster, all-optical Internet—many com- circuits.
they both map to the same logical struc- petitors are vying to create such devices. But even if the dream of an all-opti-
tures in MindNet. “It’s not perfect Large telecom equipment makers, includ- cal Internet comes to pass, another prob-
grokking,” Jensen concedes. “But it’s a ing Lucent Technologies, Agilent Tech- lem looms. So far, network designers
darn good first step.” nologies and Nortel Networks, as well as have found ingenious ways to pack more
MindNet also promises to be a pow- a number of startup companies, are and more information into fiber optics,
erful tool for machine translation, Jensen developing new optical switches and both by improving the fibers and by
says. The idea is to have MindNet create devices. Their innovations include tiny using tricks like WDM. But within five to
separate conceptual webs for English and micromirrors, silicon waveguides, even 10 years, some experts fear it won’t be
another language, Spanish, for example, microscopic bubbles to better direct light. possible to squeeze any more data into
and then align the webs so that the Eng- But none of these fixes has the tech- existing fiber optics.
lish logical forms match their Spanish nical elegance and widespread utility of The way around this may be a type of
equivalents. MindNet then annotates photonic crystals. In Joannopoulos’ lab, photonic crystal recently created by
these matched logical forms with data photonic crystals are providing the means Joannopoulos’ group: a “perfect mirror”
from the English-Spanish translation to create optical circuits and other small, that reflects specific wavelengths of light
memory, so that translation can proceed inexpensive, low-power devices that can from every angle with extraordinary effi-
smoothly in either direction. carry, route and process data at the speed ciency. Hollow fibers lined with this
Indeed, says Jensen, who is now in the of light. “The trend is to make light do as reflector could carry up to 1,000 times
process of passing on the leadership of the many things as possible,” Joannopoulos more data than current fiber optics—
group to the younger generation, Mind- says. “You may not replace electronics offering a solution when glass fibers reach
Net seems to tie together everything completely, but you want to make light do their limits. And because it doesn’t absorb
they’ve been doing for the past nine years: as much as you can.” and scatter light like glass, the invention
“All we see is doors opening. We don’t see Conceived in the late 1980s, photon- may also eliminate the expensive signal
any closing!” —M. Mitchell Waldrop ic crystals are to photons what semicon- amplifiers needed every 60 to 80 kilome-

108 TECHNOLOGY REVIEW January/February 2001


MIT’s John Joannopoulos is using photonic
crystals, like this two-dimensional version he’s
holding, to better control the flow of light.

Photograph by J O H N S O A R E S TECHNOLOGY REVIEW January/February 2001 109


With his “aspect-oriented” approach, Xerox
PARC’s Gregor Kiczales is making computer
programs easier to write and maintain.

ters in today’s optical networks (see “Blaz- GREGOR KICZALES a rule, such as: “When adding a new func-
ing Data,” TR November/December 2000). tion to this application, always put a trace
Joannopoulos is now exploring the Untangling Code statement in.” Of course, the rule works
theoretical limits of photonic crystals. ity software engineers. With the only if people remember to follow it.
How much smaller can devices be made,
and how can they be integrated into opti-
cal chips for use in telecommunications
P touch of a button, their programs
let us make global fixes in a long
text, say, or a spreadsheet, yet program-
Other crosscutting capabilities include
security and synchronization—the abili-
ty to make sure that two users don’t try to
mers often need to correct their own work access the same data at the same time.
Others Untangling Code one tedious line at a time. That irony Both require programmers to write the
isn’t lost on Gregor Kiczales, principal same functionality into many different
Organization Project
scientist at Xerox’s Palo Alto Research areas of the application. Even a modest-
Mehmet Aksit Composition filters Center (PARC) and professor at the Uni- sized application can easily present 100
(University of Twente,
the Netherlands) versity of British Columbia in Vancou- crosscutting issues.
ver—and he has a fix in mind. Kiczales Programmers try to track these
Karl Lieberherr Adaptive programming
(Northeastern University) champions what he calls “aspect-oriented instances of repetition, so that when a
programming,” a technique that will allow capability needs to be changed or upgrad-
IBM Research HyperJ system for Java
(Yorktown Heights, N.Y.) programming
software writers to make the same kinds ed, it can be done uniformly throughout
of shortcuts that those of us in other pro- the program. But keeping track of cross-
Mira Mezini Enhancing modularity and fessions have been making for years. cutting concerns is an error-prone process.
(Univ. of Siegen, Germany) reusability of A-O software
One such “crosscutting” capability is Forget to upgrade just a few of these
logging—the ability to trace and record instances, and your code starts collecting
and, perhaps, ultrafast optical computers? every operation the application performs. bugs. “We’re forced to keep track of every-
Says Joannopoulos: “Once you start being Since any given command might touch thing in our heads,” says Kiczales.
able to play with light, a whole new world down on functionally unrelated areas of Kiczales’ proposed solution is to cre-
opens up.” —Erika Jonietz the code, programmers now must make ate a new category within a program-

110 TECHNOLOGY REVIEW January/February 2001 Photograph by M A R K G I L B E R T


ming language called an “aspect.” Aspects hope to make aspects part of the ver- automate the design and manufacture of
allow programmers to write, view and nacular of programming languages. robotics by deploying computers to con-
edit a crosscutting concern as a separate “AspectJ lets programmers work more ceive, test and even build the configura-
entity. Once the programmer is happy quickly and at a higher design level,” says tions of each robotic system: in short, to
with it, a single keystroke will weave the Kiczales. “We’ve learned that crosscutting use robots to build robots. Last year, in a
aspect into the code wherever it is need- concerns are actually not hard to work cramped lab at Brandeis University in
ed. It’s a smart, intuitive, neat solution to with—once you have the proper pro- Waltham, Mass., Jordan Pollack demon-
an old problem. And what’s good for gramming support.” —Claire Tristram strated how this automated robotic design
programmers is good for the rest of us: and manufacturing might work.
Widespread adoption of aspects holds Pollack, an associate professor of com-
out the promise of less buggy upgrades, JORDAN POLLACK puter science, together with postdoc Hod
shorter product cycles and, ultimately, Lipson, directed a computer to design a
better and less expensive software. Robot Design moving creature using a limited set of
The idea of aspects has been around obot builders make a convinc- simple parts: plastic rods, ball joints, small
for many years and with many different
names. It’s called “adaptive programming”
at Northeastern University, “subjective
R ing case that in 2001, robots are
where personal computers were
in 1980—poised to break into the mar-
motors and a “brain” (neural network).
The computer—using an algorithm
inspired by biological evolution—
programming” at IBM, “composition fil- ketplace as common corporate tools and “evolved” hundreds of generations of
tering” at the University of Twente in the ubiquitous consumer products perform- potential designs, killing off the sluggish
Netherlands and “multidimensional sep- ing life’s tedious chores. One big obstacle and refining the strong. Eventually, several
aration of concerns” elsewhere. But unlike remains: It is expensive to design and of the fastest and fittest came to life,
these other research projects, Kiczales and make robots smart enough to adapt read-
Others in Robot Design
his team at PARC have taken the concept ily to different tasks and physical envi-
out of the lab and into the real world by ronments, the way human beings do. Organization Project
incorporating the idea of aspects into a That’s the reason why robotics have, Sarcos Robots for industry, medicine,
new extension of the programming lan- so far, found a commercial niche only in (Salt Lake City, Utah) Hollywood
guage Java. The beta version of this exten- simple and highly repetitive jobs, such as iRobot Household communications
sion (called AspectJ) is available for free at working on an automotive assembly line, (Somerville, Mass.) robot
www.aspectj.org, and Kiczales plans to or mass-producing identical items, such
Humanoid Interaction Lab Interactions between humanoid
make release 1.0 ready by June. “Major as toys. The challenge for builders of (Tsukuba, Japan) robots and humans
changes in programming methodology robots is to build more complexity into
can take 30 years to gain widespread accep- them without the huge investment of MIT Artificial Intelligence Machine learning, robot legs,
Lab (Cambridge, Mass.) faces
tance,” he says. Making aspects an exten- custom-tailoring each robot for a differ-
sion to an existing standard should, he pre- ent task. Robotics Institute Mobile robots and face
(Carnegie Mellon) recognition
dicts, “cut the cycle by 15 or 20 years.” One promising approach is to fully
While Kiczales admits the tools are
still a little raw, there are nevertheless
about 500 users of AspectJ today—most
of them finding existing tools inadequate
for creating long, complicated programs
in Java. Some have already found Aspec-
tJ so solid that they’ve used it in produc-
tion. One of these is Checkfree.com, a
company that makes software for auto-
matic bill payment. Checkfree sells both
C++ and Java versions of the software.
Rich Price, senior engineer, estimates
that AspectJ allowed his team to imple-
ment an important crosscutting capa-
bility in the Java version in four pro-
grammer-hours, whereas the C++ team,
with no aspect-oriented programming
tools at their disposal, took two pro-
grammer-weeks to do the same thing.
Using aspects, he says, “I make one
change, in one place, and it gets woven in
where it needs to be. I love that.”
By folding their ideas into a practical
Java extension, Kiczales and his team Brandeis’ Jordan Pollack with some of his “creatures” designed by robots.

Photograph by J O H N G O O D M A N TECHNOLOGY REVIEW January/February 2001 111


manufactured in a rapid-prototyping hold, though, they will need their own ver- approaches involved, particularly silicon
machine. Pollack and Lipson snapped sion of Moore’s Law: becoming dramat- micromachining, are so expensive that
on the motors, and the creatures moved. ically more affordable and powerful over experts in the field question whether
“I think the important point of our time. In spite of intriguing experiments products relying on these techniques
coevolutionary design and automated such as Pollack’s, designing even relative- could ever be economical to manufacture.
manufacturing for robotics is to get small- ly simple robots is a painstaking task. In Quake’s group is one of several now
quantity production to be economical,” Japan, for example, Honda has spent over working their way around these obstacles.
Pollack says. He predicts that the evolu- 14 years building a humanoid robot able Last spring, the team unveiled a set of
tionary approach to robot building could to walk, open a door and navigate stairs. microfabricated valves and pumps—a
lead to the first cheap industrial robots in A walk around Pollack’s lab suggests, critical first step in developing technolo-
five to 10 years. “If we are successful, we perhaps, a better way to design robots. On gy general enough to work for any
could see an industry within a decade a workbench sits one example of his microfluidics application. And to make
which makes low-quantity custom ma- computer-designed and computer-build- microfluidic devices cheaper, Quake and
chinery worth more than it costs to make.” able machines; it moves eerily like an others are casting them out of soft silicone
For now, Pollack’s “automated” inchworm. Pollack trims excess plastic rubber in reusable molds, using a tech-
process still takes plenty of human inter- from a newly fabricated plastic-rod nique called “soft lithography.” The
vention and money: Pollack and his col- machine, oblivious to the shavings col- potential payoff of these advances is huge:
league wrote the computer program and lecting on his shirt and around his chair. mass-produced, disposable microfluidic
spent $50,000 on the human-built fabri- In a few years Pollack may well evolve a chips that make possible everything from
cating device. Still, the team’s advance, cheap robot able to sweep those shavings drug discovery on a massive scale to at-
reported last August in the journal off the floor. —David Talbot home tests for common infections.
Nature, garnered wide publicity. “The Because microfluidics is so promising
importance is symbolic,” says Hans and yet so technically frustrating, expec-
Moravec, principal research scientist at STEPHEN QUAKE tation and hype have sometimes out-
the Robotics Institute at Carnegie Mellon paced the development of viable tech-
University in Pittsburgh. “You have sys- Microfluidics nology. Yet Quake and his group have
tems that develop robots out of thin air, he forces of physics move oceans, consistently turned out elegant devices
not by humans. In the future, there will be
real robots designed that way.”
Pollack’s design and manufacturing
T mountains and galaxies. But
applied physicist Stephen Quake
uses them to manipulate things on a vast-
that actually work. First was a microscale
DNA analyzer that operates faster and on
different principles than the conven-
methods have plenty of competition. ly reduced scale: tiny volumes of fluids tional, full-sized version, then a miniature
Academic and industrial labs around the thousands of times smaller than a dew- cell sorter and most recently, those valves
world are busy building new generations drop. Microfluidics, as Quake’s field is and pumps, described last April in the
called, is a promising new branch of journal Science. All this while regularly
Others in Microfluidics biotechnology. The idea is that once you publishing important findings on the
Organization Project master fluids at the microscale, you can basic physics of biological molecules.
automate key experiments for genomics If Quake seems adept at straddling
Aclara BioSciences Genomics and drug screening
(Mountain View, Calif.) and pharmaceutical development, per- fields—in this case science and technol-
form instant diagnostic tests, even build ogy—perhaps it’s because that’s exactly
Caliper Technologies DNA, RNA and protein assays
(Mountain View, Calif.)
implantable drug-delivery devices—all the sort of challenge he has long craved.
on mass-produced chips. It’s a vision so Even as an undergraduate at Stanford
Cepheid DNA analysis compelling that many industry observers University, where he earned bachelor’s
(Sunnyvale, Calif.)
predict microfluidics will do for biotech and master’s degrees simultaneously in
Micronics Diagnostics and chemical what the transistor did for electronics. only four years, Quake worried that
(Redmond,Wash.) analysis
Quake’s 11-person lab at Caltech is physics was “somewhat finished” as an
TECAN Drug discovery not the only outfit bent on realizing this experimental science, that it was hard to
(Hombrechtikon, Switz.)
vision. Over the past decade or so, scores find the field’s frontiers. A pioneer at
of researchers have set out to build heart, Quake started looking to tackle
of robots. Within this decade, experts microscale devices for many of the basic questions that lay at the boundaries
predict a steady evolution in commercial processes of biological research, from between disciplines. As he recalls: “It was
utility robots: robots that can clean floors sample mixing to DNA sequencing. But completely obvious, even to an outsider,
and pick up things. “There will be a mass many of those groups have run into road- that biology was going through this peri-
market for robots,” suggests George blocks in developing technology that can od of incredible growth and intellectual
Bekey, founder of the robotics lab at the be generalized to a broad range of appli- excitement, and there were going to be big
University of Southern California in Los cations and would allow several func- questions asked and answered, and the
Angeles. “This next decade will be the tions—such as sample preparation, DNA frontiers were advancing at a tremen-
decade of the robot.” extraction and detection of a gene muta- dous rate in all directions.”
Before robots reach out into the tion—to be integrated on a single chip. After Quake finished his doctorate in
everyday world of business and the house- Moreover, some of the manufacturing theoretical physics at Oxford University,

112 TECHNOLOGY REVIEW January/February 2001


Caltech’s Stephen Quake has set his
sights on the microscale, building
tiny disposable devices that could
revolutionize biotechnology.

he went back to Stanford as a fellow TR went to press, the company was plan- interest him the most. And though he has
working on the physics of DNA. When ning to deliver its first microfluidic built quite a reputation as a technologist,
Caltech’s applied physics department devices to selected university researchers he hopes soon to focus more of his atten-
hired him in 1996, Quake says, “it was an and industry partners by the end of 2000, tion on some of the most pressing ques-
experiment for them”—he was the first and was hoping for a commercial release tions in basic biology: How do the pro-
faculty member in the department with by the end of this year or early 2002. The teins that control gene expression work?
a biological bent. So far, the experiment competition will be intense. Several star- How can you do studies that cut across
seems to be going smoothly; this past tups and even electronics giants like the entire genome? “Now that we’ve got
summer, at only 31, Quake got tenure. Hewlett-Packard and Motorola are get- some pretty neat tools,” Quake says, “we’re
Quake’s inventions are also thriving ting in on the game. But to date, only one going to try and do some science with
in industry, through a startup called of Mycometrix’s competitors has brought them.” Quake’s ability to work in areas
Mycometrix. Founded in 1999 by Quake, a microfluidic product to market. from basic research to hot commercial
two of his college classmates and a con- Although Quake’s work is rapidly markets make him a prototypical inno-
sultant, the South San Francisco-based flowing into the commercial market- vator. And the same versatility makes
company has licensed all of Quake’s place, it’s still the very early stages of sci- microfluidics a field to pay close attention
microfluidics patents from Caltech. When ence and technology development that to in the next few years.—Rebecca Zacks

Photograph by M I C H A E L G R E C C O TECHNOLOGY REVIEW January/February 2001 113